- Email: [email protected]
Engineered Metal Binding Sites on Green Fluorescence Protein
Biochemical and Biophysical Research Communications 268, 462– 465 (2000) doi:10.1006/bbrc.1999.1244, available online at http://www.idealibrary.com on
Engineered Metal Binding Sites on Green Fluorescence Protein Todd A. Richmond, 1 Terry T. Takahashi, Riti Shimkhada, and Jennifer Bernsdorf Joint Science Department, Claremont McKenna, Pitzer, and Scripps Colleges, 925 North Mills Avenue, Claremont, California 91711
Received July 2, 1999
The ability to assay a variety of metals by noninvasive methods has applications in both biomedical and environmental research. Green fluorescent protein (GFP) is a protein isolated from coelenterates that exhibits spontaneous fluorescence. GFP does not require any exogenous cofactors for fluorescence, and can be easily appended to other proteins at the DNA level, producing a fluorescence-labeled target protein in vivo. Metals in close proximity to chromophores are known to quench fluorescence in a distance-dependent fashion. Potential metal binding sites on the surface of GFP have been identified and mutant proteins have been designed, created, and characterized. These metal-binding mutants of GFP exhibit fluorescence quenching at lower transition metal ion concentrations than those of the wild-type protein. These GFP mutants represent a new class of protein-based metal sensors. © 2000 Academic Press Key Words: green fluorescent protein; metal binding; fluorescence quenching; mutagenesis.
Green fluorescent protein (GFP, Fig. 1) is a spontaneously fluorescent protein isolated from coelenterates such as the Pacific jellyfish, Aequoria victoria (1). GFP is comprised of 238 amino acids, and has a major absorbance/excitation peak at 395 nm with a minor peak at 475 nm (extinction coefficients of roughly 30,000 and 7,000 M ⫺1 cm ⫺1, respectively), while the emission peak is at 508 nm (2). The structure of GFP has been solved, and the fluorophore originates from an autocatalytic cyclization of an internal ser-tyr-gly sequence (space-filling portion in the center of the protein, Fig. 1) (3). The molecular cloning of GFP cDNA (4) and the demonstration that GFP can be expressed as a functional transgene (5) has enabled experiments in new areas of cell, developmental and molecular biology. Fluorescent GFP has been expressed in bacteria (5), yeast (6), plants (7), Drosophila (8), and in mam1
To whom correspondence should be addressed. Fax: (909) 6218588. E-mail: [email protected]. 0006-291X/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.
malian cells (9). GFP tolerates N- and C-terminal fusion to a broad variety of proteins, many of which have been shown to retain function (10). Thus GFP can be easily appended to other proteins at the DNA level, producing a fluorescence-labeled target protein in vivo. This method does not require posttranslational or chemical modification of the system to be assayed. The enormous versatility of GFP as a noninvasive marker in living cells allows for a wide variety of applications such as tracing cell lineage, reporting gene expression, and as a potential measure of protein-protein interactions (11). Our studies seek to extend these uses by giving GFP and GFP-chimeras the ability to monitor metal ion presence and concentration in biological systems. The publication of crystallographic structural data allows for the design of directed modifications that confer selective metal binding to GFP. As the presence of metal ions in close proximity to a chromophore can result in fluorescence quenching (12), this introduced functionality allows the protein to be used as a metal sensor for noninvasive studies in biological and environmental systems (Fig. 2). These mutants represent an entirely new class of GFP’s that can report metal ion concentrations without the use of any exogenous modification reagents. MATERIALS AND METHODS Molecular modeling. Coordinates for GFP were obtained from the Brookhaven database. Initial visualization studies were performed using WebLab Viewer Lite (MSI). “Virtual Mutagenesis” was accomplished using HyperChem 5.1 (Hyper). Geometry optimization of mutated structures was determined by AMBER force field calculations, with all residues within 6 Å of the introduced mutation allowed to move (all other residues frozen). Metal binding geometries analyzed by semi-empirical calculations using ZINDO. Site-directed mutagenesis. Specific mutations were introduced into the 10C-GFP plasmid using the Quick-Change kit from Stratagene. Oligos coding for the desired mutations (along with their complement) were as follows: S147H, 5⬘-GGACACAAATTGGAATACAACTATAACCATCACAATGTATACATCATGGCAGACAAA-3⬘; Q204H, 5⬘-CCATTACCTGTCCTATCATTCTGCCCTTTCGAAAG-3⬘; S202D/E, 5⬘-TTTTACCAGACAACCATTACCTGGA(G/C)TATCATTCTGCCCTTTCGAAAG-3⬘; F223D/E, 5⬘-CACATGGTCCTTCTTGAG-
462
Vol. 268, No. 2, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2.
Overall scheme for design of metal-sensing GFP.
render it most useful for our studies (14). Initial molecular modeling experiments were performed using WebLab Viewer to locate pairs of residues on the sur-
FIG. 1. Overall structure of green fluorescence protein. Chromophore portion is shown in space filling representation in the interior of the protein. Coordinates from Tsien et al. (3). Modeled using WebLab Viewer (MSI).
GA(G/C)CATTGTCGACGACCCTAATGT-3⬘. The resulting plasmids were sequenced by traditional automated di-deoxy techniques. Protein expression and purification. The 10C mutant was obtained in a plasmid based on pET (T7 expression system), and also included a cleavable His-tag on the n-terminus. This His-tag was left intact during all experiments. Plasmids were transformed into competent BL21DE3 bacterial strains and expressed under typical conditions. Crude lysate was obtained and mutant GFP was purified by affinity chromatography on a nickel-agarose column (final purity ⬎95%, data not shown). Purified protein was dialyzed into either phosphate or Tris buffers in preparation for spectroscopy experiments. Quenching experiments. The 10C GFP mutants have an excitation maximum at 513 nm and emission maximum at 527 nm. A Spectrafluor II plate reader with suitable cutoff filters (ex: 480 nm, em: 525 nm) was used for all quenching experiments.
RESULTS AND DISCUSSION A mutant of GFP was used as the starting point for design of the metal binding variants (13). This variant (“10C”, T203Y/S65G/V68L/S72A) has properties which
FIG. 3. Computational mutant (10C-S147H/Q204H) with modeled Cu 2⫹ bound. Mutations created using HyperChem 5.1 Pro (Hyper). The histidine mutations were introduced and the modified residue geometry optimized using the AMBER force field. Following local minimization, all residues within 6 Å of the mutation were also geometry optimized in a similar fashion. Further modeling on the metal binding site was performed using semi-empirical methods (ZINDO).
463
Vol. 268, No. 2, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 4. Fluorescence quenching of GFP by Cu 2⫹. In a typical experiment, purified protein (WT [10C], HHD [10C-S147H/Q204H/S202D], HHE [S147H/Q204H/F223E], high nanomolar concentrations) is added to individual wells on a 96-well plate. Different concentrations of Cu 2⫹, Ni 2⫹, or Co 2⫹ were also added (Ni 2⫹ and Co 2⫹ data not shown). Fluorescence was determined using a plate reader with excitation at 485 nm and emission at 520 nm. The quenching is calculated by determining % fluorescence, which is the difference of the fluorescence at a given metal concentration divided by the fluorescence with no metal present (corrected for protein concentration differences). The Cu 2⫹ results indicate that the HHD and HHE mutants are quenched at metal concentration approximately 10 4 less than the 10C (“wild-type”) protein.
face of the protein that satisfied three criteria: sidechain on the “outside” of the protein, sidechain in close proximity to the chromophore (quenching is distance dependent), sidechains in close proximity to each other. These initial studies indicated two desirable pairs as candidates for mutation: Ser147/Gln204 and Glu95/Gln184. With the identification of initial candidates, Hyperchem was used to create computational mutants (10C-S147H/Q204H and 10C-E95H/Q184H, Fig. 3). These calculations indicated that metal–ligand bonds could be in the range of 2–3 A, which would provide an acceptable binding site. The desired mutants were introduced by standard methods (15), and were verified by sequencing. The 95/184 mutant did not express well, so the 147/204 mutant was chosen for further study. UV-Vis spectroscopy experiments were performed with purified protein (wild-type 10C and 10C-S147H/Q204H) and NiCl 2 (data not shown). The absorbance at 513 nm is due to the chromaphore. Increasing the concentration of Ni 2⫹ leads to a decrease in the absorbance at 513 nm with concomitant increase in absorbance at 400 nm and 700 nm (broad). This behavior is seen with both the wildtype 10C protein and the mutant. A second round of molecular modeling focussed on adding a nascent third ligand (a carboxylate) to the putative metal binding site. Results showed that conversion of either residue 202 or 223 to aspartic or glutamic acid could result in a third potential protein metal ligand. A second round of mutagenesis was performed, with the mutations verified by sequencing. The proteins 10C-S147H/Q204H/S202D (HHD) and 10C-
S147H/Q204H/F223E (HHE) were chosen for further quenching experiments. Fluorescence quenching was assayed using a fluorescence plate reader. Purified protein (wild-type 10C, and mutants 10C-S147H/Q204H/S202D (HHD) and 10C-S147H/Q204H/F223E (HHE) were added to individual wells in a 96-well plate. Different concentrations of Cu 2⫹, Ni 2⫹, or Co 2⫹ were added. Results from the Cu 2⫹ experiments are shown in Fig. 4 (other data not shown). The mutants are quenched at a much lower metal concentration than is the wild-type 10C protein. This indicates that the mutant proteins are able to bind metal, and that metal ion is close enough to quench the fluorophore, presumable by energy transfer. The data indicates an approximate binding constant in the low micromolar range. Further studies are underway to assay different metal ions, and to increase the affinity of the metal binding site. These mutants represent the first generation of a new class of metal sensors based on green fluorescence protein. ACKNOWLEDGMENTS The GFP-containing plasmids were obtained from Dr. Roger Tsien at the University of California, San Diego. Flourescence Spectrometer provided by Dr. Cynthia Selassie at Pomona College. This work was supported in part by grants from the W. M. Keck Foundation (summer support for T.T.T., R.S., J.B., and T.A.R.).
REFERENCES
464
1. Morin, J., and Hastings, J. (1971) J. Cell Physiol. 77, 318 –328. 2. Kahana, J., and Silver, P. in Current Protocols in Molecular
Vol. 268, No. 2, 2000
3. 4. 5. 6. 7.
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Biology (F. Ausabel, et al., Eds.), pp. 9.7.22–9.7.28. Green and Wiley, NY. Ormo, M., Cubitt, A., Kallio, L., Gross, L., Tsien, R., and Remington, S. (1996) Science 273, 1392–1395. Yang, F., Moss, L., and Phillips, G. (1996) Nature/Biotech 14, 1246 –1251. Cody, C., Prasher, D., Westler, W., Pendergast, F., and Ward, W. (1993) Biochemistry 32, 1212–1218. Prasher, D., Eckenrode, V., Ward, W., Prendergast, F., and Cormier, M. (1992) Gene 111, 229 –233. Chalfie, M., Tu, Y., Euskirchen, G., Ward, W., and Prasher, D. (1994) Science 263, 802– 805.
8. Kahana, J., Schapp, B., and Silver, P. (1995) Proc. Natl. Acad. Sci. USA 92, 9707–9711. 9. Casper, S., and Holt, C. (1996) Gene 173, 69 –73. 10. Wang, S., and Hazelrigg, T. (1994) Nature 369, 400 – 403. 11. Ludin, B., Doll, T., Meill, R., Kaech, S., and Matus, A. (1996) Gene 173, 107–111. 12. Cubit, A., Heim, R., Adams, S., Boyd, A., Gross, L., and Tsien, R. (1995) TIBS 20, 448 – 455. 13. Mitra, R., Silva, C., and Youvan, D. (1996) Gene 173, 13–17. 14. Chen, R. (1986) Anal. Let. 19, 963–977. 15. Heim, R., Prasher, D., and Tsien, R. (1994) Proc. Natl. Acad. Sci. USA 91, 12501–12504.
465
Engineered Metal Binding Sites on Green Fluorescence Protein Todd A. Richmond, 1 Terry T. Takahashi, Riti Shimkhada, and Jennifer Bernsdorf Joint Science Department, Claremont McKenna, Pitzer, and Scripps Colleges, 925 North Mills Avenue, Claremont, California 91711
Received July 2, 1999
The ability to assay a variety of metals by noninvasive methods has applications in both biomedical and environmental research. Green fluorescent protein (GFP) is a protein isolated from coelenterates that exhibits spontaneous fluorescence. GFP does not require any exogenous cofactors for fluorescence, and can be easily appended to other proteins at the DNA level, producing a fluorescence-labeled target protein in vivo. Metals in close proximity to chromophores are known to quench fluorescence in a distance-dependent fashion. Potential metal binding sites on the surface of GFP have been identified and mutant proteins have been designed, created, and characterized. These metal-binding mutants of GFP exhibit fluorescence quenching at lower transition metal ion concentrations than those of the wild-type protein. These GFP mutants represent a new class of protein-based metal sensors. © 2000 Academic Press Key Words: green fluorescent protein; metal binding; fluorescence quenching; mutagenesis.
Green fluorescent protein (GFP, Fig. 1) is a spontaneously fluorescent protein isolated from coelenterates such as the Pacific jellyfish, Aequoria victoria (1). GFP is comprised of 238 amino acids, and has a major absorbance/excitation peak at 395 nm with a minor peak at 475 nm (extinction coefficients of roughly 30,000 and 7,000 M ⫺1 cm ⫺1, respectively), while the emission peak is at 508 nm (2). The structure of GFP has been solved, and the fluorophore originates from an autocatalytic cyclization of an internal ser-tyr-gly sequence (space-filling portion in the center of the protein, Fig. 1) (3). The molecular cloning of GFP cDNA (4) and the demonstration that GFP can be expressed as a functional transgene (5) has enabled experiments in new areas of cell, developmental and molecular biology. Fluorescent GFP has been expressed in bacteria (5), yeast (6), plants (7), Drosophila (8), and in mam1
To whom correspondence should be addressed. Fax: (909) 6218588. E-mail: [email protected]. 0006-291X/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.
malian cells (9). GFP tolerates N- and C-terminal fusion to a broad variety of proteins, many of which have been shown to retain function (10). Thus GFP can be easily appended to other proteins at the DNA level, producing a fluorescence-labeled target protein in vivo. This method does not require posttranslational or chemical modification of the system to be assayed. The enormous versatility of GFP as a noninvasive marker in living cells allows for a wide variety of applications such as tracing cell lineage, reporting gene expression, and as a potential measure of protein-protein interactions (11). Our studies seek to extend these uses by giving GFP and GFP-chimeras the ability to monitor metal ion presence and concentration in biological systems. The publication of crystallographic structural data allows for the design of directed modifications that confer selective metal binding to GFP. As the presence of metal ions in close proximity to a chromophore can result in fluorescence quenching (12), this introduced functionality allows the protein to be used as a metal sensor for noninvasive studies in biological and environmental systems (Fig. 2). These mutants represent an entirely new class of GFP’s that can report metal ion concentrations without the use of any exogenous modification reagents. MATERIALS AND METHODS Molecular modeling. Coordinates for GFP were obtained from the Brookhaven database. Initial visualization studies were performed using WebLab Viewer Lite (MSI). “Virtual Mutagenesis” was accomplished using HyperChem 5.1 (Hyper). Geometry optimization of mutated structures was determined by AMBER force field calculations, with all residues within 6 Å of the introduced mutation allowed to move (all other residues frozen). Metal binding geometries analyzed by semi-empirical calculations using ZINDO. Site-directed mutagenesis. Specific mutations were introduced into the 10C-GFP plasmid using the Quick-Change kit from Stratagene. Oligos coding for the desired mutations (along with their complement) were as follows: S147H, 5⬘-GGACACAAATTGGAATACAACTATAACCATCACAATGTATACATCATGGCAGACAAA-3⬘; Q204H, 5⬘-CCATTACCTGTCCTATCATTCTGCCCTTTCGAAAG-3⬘; S202D/E, 5⬘-TTTTACCAGACAACCATTACCTGGA(G/C)TATCATTCTGCCCTTTCGAAAG-3⬘; F223D/E, 5⬘-CACATGGTCCTTCTTGAG-
462
Vol. 268, No. 2, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2.
Overall scheme for design of metal-sensing GFP.
render it most useful for our studies (14). Initial molecular modeling experiments were performed using WebLab Viewer to locate pairs of residues on the sur-
FIG. 1. Overall structure of green fluorescence protein. Chromophore portion is shown in space filling representation in the interior of the protein. Coordinates from Tsien et al. (3). Modeled using WebLab Viewer (MSI).
GA(G/C)CATTGTCGACGACCCTAATGT-3⬘. The resulting plasmids were sequenced by traditional automated di-deoxy techniques. Protein expression and purification. The 10C mutant was obtained in a plasmid based on pET (T7 expression system), and also included a cleavable His-tag on the n-terminus. This His-tag was left intact during all experiments. Plasmids were transformed into competent BL21DE3 bacterial strains and expressed under typical conditions. Crude lysate was obtained and mutant GFP was purified by affinity chromatography on a nickel-agarose column (final purity ⬎95%, data not shown). Purified protein was dialyzed into either phosphate or Tris buffers in preparation for spectroscopy experiments. Quenching experiments. The 10C GFP mutants have an excitation maximum at 513 nm and emission maximum at 527 nm. A Spectrafluor II plate reader with suitable cutoff filters (ex: 480 nm, em: 525 nm) was used for all quenching experiments.
RESULTS AND DISCUSSION A mutant of GFP was used as the starting point for design of the metal binding variants (13). This variant (“10C”, T203Y/S65G/V68L/S72A) has properties which
FIG. 3. Computational mutant (10C-S147H/Q204H) with modeled Cu 2⫹ bound. Mutations created using HyperChem 5.1 Pro (Hyper). The histidine mutations were introduced and the modified residue geometry optimized using the AMBER force field. Following local minimization, all residues within 6 Å of the mutation were also geometry optimized in a similar fashion. Further modeling on the metal binding site was performed using semi-empirical methods (ZINDO).
463
Vol. 268, No. 2, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 4. Fluorescence quenching of GFP by Cu 2⫹. In a typical experiment, purified protein (WT [10C], HHD [10C-S147H/Q204H/S202D], HHE [S147H/Q204H/F223E], high nanomolar concentrations) is added to individual wells on a 96-well plate. Different concentrations of Cu 2⫹, Ni 2⫹, or Co 2⫹ were also added (Ni 2⫹ and Co 2⫹ data not shown). Fluorescence was determined using a plate reader with excitation at 485 nm and emission at 520 nm. The quenching is calculated by determining % fluorescence, which is the difference of the fluorescence at a given metal concentration divided by the fluorescence with no metal present (corrected for protein concentration differences). The Cu 2⫹ results indicate that the HHD and HHE mutants are quenched at metal concentration approximately 10 4 less than the 10C (“wild-type”) protein.
face of the protein that satisfied three criteria: sidechain on the “outside” of the protein, sidechain in close proximity to the chromophore (quenching is distance dependent), sidechains in close proximity to each other. These initial studies indicated two desirable pairs as candidates for mutation: Ser147/Gln204 and Glu95/Gln184. With the identification of initial candidates, Hyperchem was used to create computational mutants (10C-S147H/Q204H and 10C-E95H/Q184H, Fig. 3). These calculations indicated that metal–ligand bonds could be in the range of 2–3 A, which would provide an acceptable binding site. The desired mutants were introduced by standard methods (15), and were verified by sequencing. The 95/184 mutant did not express well, so the 147/204 mutant was chosen for further study. UV-Vis spectroscopy experiments were performed with purified protein (wild-type 10C and 10C-S147H/Q204H) and NiCl 2 (data not shown). The absorbance at 513 nm is due to the chromaphore. Increasing the concentration of Ni 2⫹ leads to a decrease in the absorbance at 513 nm with concomitant increase in absorbance at 400 nm and 700 nm (broad). This behavior is seen with both the wildtype 10C protein and the mutant. A second round of molecular modeling focussed on adding a nascent third ligand (a carboxylate) to the putative metal binding site. Results showed that conversion of either residue 202 or 223 to aspartic or glutamic acid could result in a third potential protein metal ligand. A second round of mutagenesis was performed, with the mutations verified by sequencing. The proteins 10C-S147H/Q204H/S202D (HHD) and 10C-
S147H/Q204H/F223E (HHE) were chosen for further quenching experiments. Fluorescence quenching was assayed using a fluorescence plate reader. Purified protein (wild-type 10C, and mutants 10C-S147H/Q204H/S202D (HHD) and 10C-S147H/Q204H/F223E (HHE) were added to individual wells in a 96-well plate. Different concentrations of Cu 2⫹, Ni 2⫹, or Co 2⫹ were added. Results from the Cu 2⫹ experiments are shown in Fig. 4 (other data not shown). The mutants are quenched at a much lower metal concentration than is the wild-type 10C protein. This indicates that the mutant proteins are able to bind metal, and that metal ion is close enough to quench the fluorophore, presumable by energy transfer. The data indicates an approximate binding constant in the low micromolar range. Further studies are underway to assay different metal ions, and to increase the affinity of the metal binding site. These mutants represent the first generation of a new class of metal sensors based on green fluorescence protein. ACKNOWLEDGMENTS The GFP-containing plasmids were obtained from Dr. Roger Tsien at the University of California, San Diego. Flourescence Spectrometer provided by Dr. Cynthia Selassie at Pomona College. This work was supported in part by grants from the W. M. Keck Foundation (summer support for T.T.T., R.S., J.B., and T.A.R.).
REFERENCES
464
1. Morin, J., and Hastings, J. (1971) J. Cell Physiol. 77, 318 –328. 2. Kahana, J., and Silver, P. in Current Protocols in Molecular
Vol. 268, No. 2, 2000
3. 4. 5. 6. 7.
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Biology (F. Ausabel, et al., Eds.), pp. 9.7.22–9.7.28. Green and Wiley, NY. Ormo, M., Cubitt, A., Kallio, L., Gross, L., Tsien, R., and Remington, S. (1996) Science 273, 1392–1395. Yang, F., Moss, L., and Phillips, G. (1996) Nature/Biotech 14, 1246 –1251. Cody, C., Prasher, D., Westler, W., Pendergast, F., and Ward, W. (1993) Biochemistry 32, 1212–1218. Prasher, D., Eckenrode, V., Ward, W., Prendergast, F., and Cormier, M. (1992) Gene 111, 229 –233. Chalfie, M., Tu, Y., Euskirchen, G., Ward, W., and Prasher, D. (1994) Science 263, 802– 805.
8. Kahana, J., Schapp, B., and Silver, P. (1995) Proc. Natl. Acad. Sci. USA 92, 9707–9711. 9. Casper, S., and Holt, C. (1996) Gene 173, 69 –73. 10. Wang, S., and Hazelrigg, T. (1994) Nature 369, 400 – 403. 11. Ludin, B., Doll, T., Meill, R., Kaech, S., and Matus, A. (1996) Gene 173, 107–111. 12. Cubit, A., Heim, R., Adams, S., Boyd, A., Gross, L., and Tsien, R. (1995) TIBS 20, 448 – 455. 13. Mitra, R., Silva, C., and Youvan, D. (1996) Gene 173, 13–17. 14. Chen, R. (1986) Anal. Let. 19, 963–977. 15. Heim, R., Prasher, D., and Tsien, R. (1994) Proc. Natl. Acad. Sci. USA 91, 12501–12504.
465